Workshop on Polar and Global Climate Modeling: Connection and Interplay

June 14-16, 2006
International Arctic Research Center
University of Alaska Fairbanks, Alaska, USA

[HOME] [ANNOUNCEMENT] [PARTICIPANTS] [ABSTRACTS] [AGENDA] [Printable Agenda - PDF]

Abstracts

Atmospheric heat transport and polar amplification

Vladimir Alexeev
International Arctic Research Center, University of Alaska Fairbanks

Most modeling studies show climate warming to be most pronounced at high latitudes. This effect is called polar amplification (PA) and is the focus of this study. A simplifed GCM and an energy balance model (EBM) are employed to diagnose the PA pattern in a conceptual model of the climate system without any ocean- and sea ice feedbacks. The PA pattern seems to be connected in this case to the sensitivity of the atmospheric heat transport to large scale surface temperature perturbations. The PA is also analyzed in terms of system's linearized dynamics. The PA is seen to arise as the excitation of the least stable mode of the system by the 2xCO2 forcing in this context, because the least stable mode has a polar amplified shape. The reasons why this mode is polar amplified are studied in a series of experiments with the GCM and the EBM.

A new high-resolution unstructured grid finite-volume Arctic Ocean model (AO-FVCOM): Validation and application

Changsheng Chen¹, Gaoping Gao², Jiahua Qi¹, Andrey Proshutinsky³, Robert C. Beardsley³, Huichan Lin⁴, Geoffrey Cowles¹ and Haosheng Huang¹

¹Department of Fisheries Oceanography, School for Marine Science and Technology, University of Massachusetts-Dartmouth, 706 South Rodney French Blvd, New Bedford, MA 02748
²Department of Physical Oceanography, Ocean University of China, Qingdao, P.R. China
³Department of Physical Oceanography, Woods Hole Oceanographic Institution, Woods Hole, MA 02543
⁴Marine Extension Service, University of Georgia, Athens, GA 30602

The unstructured grid, Finite-Volume, three-dimensional (3D) primitive equations Coastal Ocean Model (FVCOM) is extended to the spherical coordinates for basin and global applications.  A mass conservative discrete method is developed to take an advantage of the flux calculation of the finite-volume method on the spherical plane. The spherical-polar stereographic projection nesting unstructured grid used at the North Pole solves the integral form of the governing equations in a spherical coordinate system without the need of either “grid rotation” or “projection” of the entire computational domain.  The spherical coordinate FVCOM is validated by the comparison with the analytical wave solution over an idealized spherical plane with and without the North Pole and application to the Arctic Ocean (AO). A high resolution AO-FVCOM (including the Greenland Sea, Hudson Bay, Baffin Bay and Canadian Arctic Archipelago) is configured. The high-resolution unstructured grid (with a horizontal resolution ranging from 1 km to 15 km) used in this model accurately resolves the complex coastal geometries, numerous islands, and narrow passages. The model has successfully simulated tidal motion in the Arctic, especially the resonant tides in the bays and the bottom-intensified internal tidal currents over the slope.  The model also resolves the current separation and thus eddy formation around islands and straits. A fully nonlinear ice model is incorporated into AO-FVCOM and used to examine the impact of tide-induced, geographically featured eddies on ice dynamics in the Arctic. With flexibility in geometric fitting and better presentation of mass conservation, AO-FVCOM provides a new tool for AO research.

Constraining snow albedo feedback with the seasonal cycle

Alex Hall
UCLA Dep't of Atmospheric and Oceanic Sciences, Los Angeles, CA  90095

Differences in simulations of climate feedbacks are sources of significant divergence in climate models' temperature response to anthropogenic forcing.  Snow albedo feedback is particularly critical for climate change prediction in heavily-populated northern hemisphere land masses.  Here we show its strength in current models exhibits a factor-of-three spread.  These large intermodel variations in feedback strength in climate change are nearly perfectly correlated with comparably large intermodel variations in feedback strength in the context of the seasonal cycle.  Moreover, the feedback strength in the real seasonal cycle can be measured and compared to simulated values. These mostly fall outside the range of the observed estimate, suggesting many models have an unrealistic snow albedo feedback in the seasonal cycle context.  Because of the tight correlation between simulated feedback strength in the seasonal cycle and climate change, eliminating the model errors in the seasonal cycle will lead directly to a reduction in the spread of feedback strength in climate change.  Though this comparison to observations may put the models in an unduly harsh light because of uncertainties in the observed estimate that are difficult to quantify, our results map out a clear strategy for targeted observation of the seasonal cycle to reduce divergence in simulations of climate sensitivity.

Modeling Tidal and Inertial Variability in Sea-Ice Drift and Deformation: Climatic Implications

W. D. Hibler, III
International Arctic Research Center, University of Alaska Fairbanks

Semidiurnal oscillations are a ubiquitous feature of Arctic and Antarctic sea-ice drift and deformation. Over much of the Arctic basin inertial and semi-diurnal tidal variability occur at similar frequencies so their periodicity alone is typically inadequate to resolve the source. To investigate the relative roles of tidal and inertial variability in the Arctic, a high-resolution pan-arctic barotropic ice ocean model with sea-ice imbedded in an upper boundary layer is constructed and numerically investigated in light of hourly observed buoy drift at several locations; and temporally dense RGPS observations of ice deformation near the pole. A 'levitated' ice ocean model--as used in almost all ice ocean models--is also examined to demonstrate some of the deleterious characteristics of this formulation. In 'levitated' models mechanical buoyancy effects of sea-ice are neglected so that convergence of ice, for example, does not affect the Ekman flux of the ocean the stress on an ocean model separate from the ice is taken to be the water drag on the underside of the ice. This artificial, but widely used, formulation may be visualized by imagining an ice cover floating above rather than in the ocean; hence the term levitated. In much of the analysis of the simulated and observed results rotary spectral techniques are utilized as the rotation sense of both sea-ice drift and deformation at the semidiurnal period provides a useful discrimanant between tidal and inertial effects over much of the Arctic Basin.

With regard to the interplay of tidal and inertial variability in sea ice drift and deformation a key result of this study is that with a properly formulated imbedded model inertial variability induced by the the wind is substantial amplified by the clockwise rotary M2 tidal forcing. Consequently the total M2-Inertial power in higher tide regions substantially exceeds the inertial power alone or the tidal forced power alone with amplification factors up to ~50 over the wind induced power alone. This feature is not however an artificial resonance and will drop back to lower tidal only forced amplitudes up to 100 times smaller in spectral power if the wind is removed. In the case of levitated sea ice models coupled to an ocean model containing an oceanic boundary layer this amplification is substantially damped and much smaller amplification factors are obtained. In addition in levitated models there is an artificial parity dependent resonance in the ice response which leads to distorted response both in parity and amplitude, especially in high amplitude tidal regions. This parity dependent response is especially notable in central Arctic deformation predictions from the levitated model which are are disagreement with observations and the embedded model in both parity and amplitude.

Masaru Inatsu*, Masahide Kimoto, and Akimasa Sumi

Center for Climate System Research, the University of Tokyo, Japan

Some aspects of global warming response in the stratosphere and the troposphere have been investigated, using the atmosphere-ocean coupled general circulation model named MIROC. Two 100-yr time integrations have performed with the atmospheric T106L56 and oceanic 1/4x1/6-L47 resolution under the current and scenario a1b climate condition. The top level of the model is 1 hPa (~40km). The years 1980-1999 and 2050-2069 are analyzed.

The sudden warming is a characteristic phenomenon in the stratosphere. The major warming, where the temperature gradient is reversed and zonal wind becomes easterly, occurs about every other year in the observation. The polar vortex surrounded by the polar night jet is gradually loosening in a normal year, while it is abruptly broken in the sudden warming year. The sudden warming is caused by the upward propagating Rossby waves with a large amplitude, which tend to decelerate the polar night jet. The MIROC remarkably reproduces such sudden warming features under the current climate. While zonal-mean temperature in the polar region is gradually warming in a normal year, it is rising by ~50K in the beginning of March in a major warming year (Figure). Estimating the global warming response, the frequency of the major warming does not change so much at least in this model, presumably because the positive effect of the weaker westerly jet in the stratosphere preferring wave intrusions and the negative effect of less Rossby waves propagating from the troposphere might be cancelled (not shown).

Figure: Zonal-mean temperature at 10 hPa in Dec to Apr in a normal year (above) and an major sudden warming year (bottom) in the current climate experiment by the MIROC. The contour interval is 2K and the red shadings is > 214K. The horizontal axis indicates days from 1 Dec.

*Corresponding author address: Dr. Masaru Inatsu, Center for Climate System Research, the University of Tokyo, General Research Bldg., 5-1-5 Kashiwanoha, Kashiwa, Chiba 2778586 Japan. E-mail: inaz @ ccsr.u-tokyo.ac.jp

Effect of summertime wind conditions on lateral and bottom melting in the central Arctic

Jun Inoue and Takashi Kikuchi
Institute of Observational Research for Global Change (IORGC), JAMSTEC
2-15 Natsushima-cho, Yokosuka, 237-0061, Japan, jun.inoue@jamstec.go.jp

To understand a role of wind conditions on summertime surface ocean system, the ocean, ice, and atmosphere in the central Arctic Ocean were observed using drifting buoys, one in 2002 under the stormy condition, one in 2003 under the calm condition. Although the ice concentration near the North Pole was the same in 2002 and 2003 during early summer, the heat used in bottom melting in 2003 was almost half of that in 2002. To obtain the total heat input into the upper ocean, heat used in lateral melting was additionally derived from a time series of ice concentration in 2002. Assuming the same heat input into the upper ocean, the heat used in lateral and bottom melting was estimated and compared between the years. It is thought that the warm fresh water embedded within the ice cover was mixed downward during the frequently stormy mid-summer of 2002, enhancing bottom melting. By contrast, the warm water in 2003 tended to be used for lateral melting due to the relatively calm conditions, suggesting that a continuously weak wind is favourable to decrease the ice cover during summer. A simple calculation of evolution of ice cover reveals that the difference in ice concentration during August between 2002 and 2003 reached 10 %, which is consistent with the satellite-derived ice concentration.

Arctic climate dynamics in the context of global warming: an interplay of the change and variability in AOGCM simulations

Vladimir Kattsov, Petr Sporyshev and Tatyana Pavlova

Voeikov Main Geophysical Observatory, St.Petersburg, Russia

The ability of AOGCMs to reproduce the current climate is an important but not sufficient prerequisite for credible climate predictions. Another basic target for the models is simulating the evolution of the climate system through the period of instrumental observations. For the Arctic, with its scarcity of observations and both spatial and temporal high variability of the climate, an evaluation of 20th century simulations faces certain challenges. Of particular interest are the two warming events observed in the first half of the 20th century and in the past several decades. Presuming the different nature of the two events – the former being a manifestation of the natural variability, while the latter, at least partly, being a result of the anthropogenic forcing – the two requirements for simulations of the arctic surface climate through the 20th century are: (1) the ability to reproduce trends associated with external forcing, i.e. the entire 20th century and the late 20th century warming; and (2) the ability to generate unforced variability with amplitudes comparable to those observed in the first two thirds of the 20th century in the corresponding frequency bands. An opportunity to perform such evaluation is provided by a few available century-long time series of arctic surface air temperature, sea level pressure and sea ice. IPCC AOGCMs evaluated accordingly have demonstrated a diversity in their ability to satisfy the both requirements, and therefore – in their applicability in predictions of arctic climate.

Of evident interest is not only the climate change measured as a ratio of a climatic variable change to the time period over which the change happened, but also the time (e.g. the year) when the evolving difference with the baseline climate becomes a change – i.e. stably statistically significant at a pre-set level (e.g. 5%). The global picture of the timing of the stable change of surface air temperature (STC) based on an ensemble of IPCC AOGCM simulations shows an encouraging similarity with available observational data. As expected the two pronounced areas of late STC can be seen in the sub-arctic North Atlantic and in the Southern Ocean. In some simulations the STC in those regions has not been achieved in the 21st century even for the “strong” A2 scenario. Relatively late is the STC in the northern North Pacific. In contrast, relative to the baseline of 1910-1959, both the simulations and observations indicate that STC happened over the Indian Ocean as early as in the third quarter of the 20th century. While the evaluation of the AOGCMs specifically in high latitudes is not possible because of the gaps in the observations, an intercomparison of the arctic geographical patterns of STC timing obtained in simulations with different models gives some food for thought.

This study was supported by the US NSF via the IARC of the University of Alaska Fairbanks (subaward UAF05-0074 of OPP-0327664), the Russian Foundation for Basic Research (05-05-65093, 06-05-64660), and INTAS (Grant 03-51-4620). We acknowledge the international modeling groups for providing their data for analysis, the Program for Climate Model Diagnosis and Intercomparison (PCMDI) for collecting and archiving the model data, the JSC/CLIVAR Working Group on Coupled Modelling (WGCM) and their Coupled Model Intercomparison Project (CMIP) and Climate Simulation Panel for organizing the model data analysis activity, and the IPCC WG1 TSU for technical support. The IPCC Data Archive at Lawrence Livermore National Laboratory is supported by the Office of Science, U.S. Department of Energy.

Freshwater distribution, flux and residence time in the Canada Basin of the Arctic Ocean: from observations

M. Yamamoto-Kawai¹, F.A. McLaughlin¹, E.C. Carmack¹, S. Nishino², K. Shimada², and A.Y. Proshutinsky^3
1Department of Fisheries and Oceans, Institute of Ocean Sciences, Sidney, British Columbia, Canada.
²Institute of Observational Research for Global Change, Japan Agency for Marin-Earth Science and Technology, Yokosuka, Kanagawa, Japan.
³Woods Hole Oceanographic Institution, Massachusetts, USA.

The Canada Basin is the largest reservoir of freshwater stored in the Arctic Ocean. Sources of this freshwater (with respect to salinity of 34.87) are sea ice meltwater, meteoric water (precipitation and runoff), and Pacific Water flowing through the Bering Strait.  Since salinity determines density and thus stratification of water masses in high latitude regions, freshwater storage in the Canada Basin plays a key role in the global climate by influencing air-ice-ocean heat exchange and formation of North Atlantic deep water.  It is therefore desirable to understand the variability of each source of freshwater, how and where it is stored presently, in order to predict change. This is done using geochemical tracers, specifically nutrients and oxygen isotope ratio of water.  Observations in 2003 and 2004 in the Canada Basin show that meteoric water is the major source of freshwater in the top 50 m layer of the water column, whereas Pacific Water is the major source below 50 m depth.  Content of each freshwater source is integrated through the water column and multiplied by the area of the Canada Basin to estimate total volume stored in the Basin.  The mean residence time of Pacific Water is calculated using these data together with estimates of transport.  Then fluxes of meteoric water and net ice growth rate are also estimated.

High-resolution coupled ocean-atmosphere modeling for climate studies

Masahide Kimoto¹, Akimasa Sumi1, Masaru Inatsu¹ and The K-1 Japan Project Team²

¹Center for Climate System Research, University of Tokyo, 4-6-1, Komaba, Meguro-ku, Tokyo, 153-8904, Japan. E-mail: kimoto@ccsr.u-tokyo.ac.jp
²Center for Climate System Research, University of Tokyo, National Institute for Environmental Sciences, and Frontier Research Center for Global Change

A moderately high-resolution coupled ocean-atmosphere general circulation model (GCM) has been developed on the Earth Simulator for global warming projection experiments. The development team, called K-1 Japan Project Team, consists of scientists at Center for Climate System Research (CCSR) of the University of Tokyo, National Institute for Environmental Studies (NIES), and the Frontier Research Center for Global Change (FRCGC) of the Japan Agency for Marine Science and Technology. The coupled climate model is called MIROC, short for a Model for Interdisciplinary Research On Climate.

The atmospheric part adopts spectral triangular truncation at wave number 106, corresponding to about 110 km transform grid and has 56 vertical levels. The ocean GCM has 1/6 degree latitude and 1/4 degree longitude horizontal resolution and 48 vertical levels. Sea ice component shares the same grid with the ocean, and in a land submodel, each of the atmospheric Gaussian grid is divided into 2 x 2 longitude-latitude grids.

The component models of the coupled model are equipped with state-of-the-art physics packages, the detail of which is described by K-1 model developers (2004). The physics packages of the component models have been subject to complete readjustment to realize stable long-term integrations.

A lower resolution version has also been developed: T42L20 AGCM and 1x1.4deg.L44 OGCM. Both versions share the same physics package and parameters (except a few like horizontal diffusion). Results of global warming projection experiments of both model versions were submitted to the IPCC Model Output web site maintained by PCMDI and are available for the world-wide research community.

Overall performance of the two versions is not drastically different on continental scales, but regional details are much more realistically represented in the higher resolution version. A set of integrations has been conducted, in which AGCM versions with T42, T106, and T213 have been coupled with the medium and high resolution versions of the OGCM. Relative impacts of atmospheric and oceanic resolutions will be discussed.

A Labrador Sea modeling studied by a coupled sea ice-ocean circulation model

Hideaki Kitauchi¹
Frontier Research Center for Global Change, JAMSTEC, Kanagawa, Japan
Hiroyasu Hasumi
Center for Climate System Research, University of Tokyo, Chiba, Japan

The Labrador Sea is one of the most extreme ocean convection sites in the World Ocean
characterized by weak density stratification, in each wintertime, exposed to intense buoy-
ancy loss to the atmosphere and broken down by deep-reaching convection (see, for review, Marshall and Schott, 1999). The open-ocean convection in the Labrador Sea mixes the surface waters to great depth (_ 1000–2300 m during the 1990s, Lazier et al., 2002) to form the intermediate depth water mass called Labrador Sea Water (LSW). The Labrador Sea is also an important component of the thermohaline circulation, i.e., the global meridional-overturning circulation of the ocean, responsible for roughly half of the net poleward heat transport demanded of the atmosphere-ocean system (Macdonald and Wunsch, 1996). Recent rapid progress of computer sciences makes it possible to resolve mesoscale eddies (20–50 km, for example, Prater 2002) in the Labrador Sea by use of an ocean circulation model. In this circumstance we are interested in estimating the formation rate of the LSW by making use of an eddy-resolving coupled sea ice-ocean circulation model (Center for Climate System Research Ocean Component Model version 4, Hasumi, 2006) applied to the northern North Atlantic.

To this end qualitative comparison between the simulated and observed fields in the Labrador Sea is performed. The mean horizontal circulation of the surface cyclonic subpolar gyre and recirculation of the mid-depth (_ 700 m) cyclonic boundary currents (Lavender et al., 2000) are well simulated. The distribution of the eddy kinetic energy (EKE), which represents the strength of the deviation field from the mean circulations, shows good agreement with the observed results (for example, Brandt et al., 2004) over the surface in the annual average. Our model captures the observed typical vortical structures of a pair of vortices and a cluster of vortices offshore West Greenland (Prater 2002), which is only achieved by use of eddy-resolving ocean circulation models.
The vertical convection, on the other hand, needs further improvement. The mixed-layer depth, at which the potential density referenced to the sea surface becomes 0.1 kg/m3 heavier than the one at the surface, is deeper than the observed evidence (for example, Pickart et al., 2002) over the Labrador Sea. It turns out that the salinity is higher than the observed distribution in the Labrador Basin. This may be caused by the less low salinity flux from the Arctic into the East and West Greenland, suggesting that the freshwater flux from the Arctic may be partly responsible for stabilization of the density stratification in the Labrador Sea.

¹Corresponding author address: Hideaki Kitauchi 3173-25 Showa-machi, Kanazawa-ku, Yokohama, Kanagawa 236-0001, Japan E-mail: kitauchi@jamstec.go.jp

Changes in Snow Cover and Snow Water Equivalent Due to Global Warming Simulated by a 20km-mesh Global Atmospheric Model

Akio Kitoh, Masahiro Hosaka and Daisuke Nohara
Meteorological Research Institute, Tsukuba, Japan, kitoh@mri-jma.go.jp, mhosaka@mri-jma.go.jp, nohara@apcc21.net

Changes in snow due to global warming are investigated by a time-slice experiment with a 20 km-mesh atmospheric general circulation model (AGCM). The seasonal march of the snow cover in the present-day simulation is comparable to that of satellite-based observational data. Due to high horizontal resolution of the model, distributions of the simulated snow cover and snow water equivalent (SWE) reflect the detailed geographical features. At the end of the twenty-first century, due to global warming, the beginning of the snow-accumulating season (the end of the snow-melting season) will occur later (earlier) in most snow regions, and the snow cover will decrease except for very few exceptions. SWE will also decrease in wide areas, but over the cold regions (Siberia and the northern parts of North America), SWE will increase due to increases of snowfall in the coldest season. In both the change and the percentage change of the SWE, the detailed geographical features are noted.

Nobumasa Komori, Akira Kuwano-Yoshida, and Wataru Ohfuchi
Earth Simulator Center, JAMSTEC, Yokohama, Japan

Coupled atmosphere–ocean general circulation model (CGCM) is now becoming an inevitable research tool for climate variation studies. And pressing concern on the earth environment not only in the global but regional scales is one of our main motivations to pursue the possibility of developing a CGCM capable of representing “mesoscale” phenomena in a relatively explicit fashion. With these in mind, we have developed CFES, a fully parallelized code of coupled atmosphere–land–ocean–sea-ice model for the Earth Simulator (ES), which consists of AFES2 (atmospheric GCM for the ES with land-surface model MATSIRO) and OIFES (ocean–sea-ice GCM for the ES).

CFES was mainly designed to achieve efficient computational performance on the ES by using multiple-program multiple-data techniques and fully parallelized coupling schemes: In order to reduce communication cost arising from gathering/broadcasting (GB) data in the conventional coupler, we circumvent this GB process and the direct parallel transfer of the variables between AFES2 and OIFES was realized in our coupled code.

Using CFES, we started a coupled atmosphere–ocean simulation with the resolutions of T239 (about 50 km) and L48 for the atmosphere and 0.25º (about 25 km) and 54 levels for the ocean. Coupling interval is an hour, and no flux adjustment is applied. It takes about 20 hours for 1-year integration using 60 nodes (30 for AFES2 and 30 for OIFES) of the ES. In this presentation, we will cover the model description, our experimental setting, and some preliminary results with emphasis on seasonal to interannual climate variability in the polar regions.

Figure: Outputs from CFES in February 11 of the model 6th year. (left) 850-hPa specific humidity (g kg-1, in shades) and sea level pressure (in contours drawn every 4 hPa). (right) Current speed at the 54-m depth (cm s-1).

Modeling Arctic Climate: Forcing, Feedbacks and Effects on Lower Latitudes

Wieslaw Maslowski, Naval Postgraduate School, Monterey, CA, USA

The Arctic Ocean operates on three basic principles. First, it receives the heat and buoyancy fluxes from the atmosphere at the surface and from lower latitude oceans via northward advection of water mass and properties. River runoff contributes significant freshwater input locally. Second, the net heat and buoyancy sources together with dynamic wind forcing modulate the state of sea ice cover, determining variability in multi-year and first-year ice distribution, regions of net growth/melt of sea ice and the amount of total freshwater content. Third, the wind- and thermohaline-driven circulation redistributes the sea ice and water masses within the Arctic Ocean and controls their export out to the North Atlantic.  It is believed that the freshwater export from the Arctic can exert a critical role on the rate of deep-water formation in the northern North Atlantic, which in turn dominates the strength of meridional overturning circulation in the North Atlantic and the global ocean thermohaline circulation. The latter, contributes to the global heat re-distribution and climate variability at longer time scales.

Details of the operation and variability of each of the above principles are not well known from observations and they have posed great challenges to global climate models. For example, the advection of oceanic heat northward through Fram Strait still remains a challenge for most global ocean and climate models. The general tendency in low resolution models is to transport most of Atlantic Water via the Barents Sea and to have Fram Strait experience outflow to the south only. This presents a problem as most of the heat entering the Barents Sea is lost to the atmosphere before entering the central Arctic Ocean, which means that oceanic heat input to the eastern Arctic might be significantly under-represented. Similarly, the inflow of Pacific Summer Water through narrow (~100 km) Bering Strait and its circulation over the Chukchi Shelf and in the Beaufort Sea is not realistic in low resolution models, which creates problems in the western Arctic. Another challenge for global climate models is representation of narrow (10-100 km) coastal and boundary currents, which in the Arctic Ocean constitute main circulation features.

The oceanic heat, in addition to atmospheric radiative and sensible heat input, contributes to sea ice melt, which in recent years have accelerated, especially in regions coincident directly downstream of oceanic heat advection from the Pacific and Atlantic oceans. Recent reduction of the Arctic ice pack has been primarily associated with anomalies of surface air temperature and circulation over the Arctic and those in turn have been linked to the Arctic Oscillation (AO). Such studies typically assume the dominant role of external atmospheric forcing and neglect effects of processes internal to the Arctic Ocean. Especially overlooked tends to be the oceanic thermodynamic control of sea ice through the under-ice ablation and lateral melt along marginal ice zones. However, those ice-ocean interactions may act to de-correlate AO forcing, which could help explain some of the timing issues between AO/atmospheric forcing and sea ice variability.

The combined contribution from sea ice, river runoff, and net precipitation determines the freshwater content and its variability in the Arctic Ocean. The upper-ocean mixing and circulation redistributes water masses throughout the basin and out to the North Atlantic. The freshwater export through Fram Strait and the Canadian Archipelago (CAA), its rate and variability presents another big challenge to global ocean and climate models. The pathway through the CAA is usually not accounted properly due to a low model grid cell resolution, which means most of the freshwater is advected out from the Arctic Ocean via Fram Strait. However, the recirculation of Atlantic Water in the Greenland, Norwegian, and Iceland seas diffuses a significant portion of the freshwater signal as it flows along the East Greenland Current and downstream through Denmark Strait before it reaches the northern North Atlantic.  This may significantly reduce the effect of freshwater on the rate of deep-water formation and the meridional overturning circulation in the North Atlantic.

In my talk, I will overview and provide examples of the above mentioned problems related to the three principles of the operation of the Arctic Ocean. Results from a high-resolution regional coupled ice-ocean model driven with realistic atmospheric forcing will be used to suggest improvements in representation of this region in climate models.

The Rossby Centre regional model system applied to the Arctic

Markus Meier
Swedish Meteorological and Hydrological Institute

A coupled atmosphere-ice-ocean model has been developed for the Arctic based upon the Rossby Centre Atmosphere-Ocean model (RCAO). Its ice-ocean component RCO is a regional version of OCCAM (of the Ocean Circulation Climate Advanced Modelling project) coupled to the Los Alamos Sea Ice Model (CICE). The ocean model covers the central Arctic Ocean and the North Atlantic roughly to 50N. The Bering Sea is also included in order to get a realistic description of the transport variability through the Bering Strait. The horizontal resolution is 0.5 degrees or approximately 50 km in a rotated coordinate system centered over the North Pole. 59 vertical levels are utilized. The atmosphere model covers the same area as the ocean model and additionally some of the surrounding landmasses. The horizontal resolution of the atmosphere model is about 50 km as well. The atmosphere model is driven at the boundaries by 6-hourly ERA-40 fields. Sensitivity studies with modified initial conditions, ice parameters, and radiation formulations have been performed. In addition to coupled runs simulations with RCO offline using 6-hourly ERA-40 data at the surface have been carried out for 1958-2002. First results of RCAO and RCO will be shown and the added value of regional climate models for the Arctic will be discussed.

Modeling feedbacks between the water and energy budgets of the Arctic Ocean

James R. Miller and Jennifer E. Francis,
Dept. of Marine and Coastal Sciences, Rutgers University, 71 Dudley Road, New Brunswick, NJ 08901

Gary L. Russell and Yonghua Chen,
NASA/Goddard Space Flight Center, Institute for Space Studies, 2880 Broadway, New York, NY 10025

There are many complex interactions and feedbacks among the components of the climate system in the Arctic Ocean. Global climate models indicate that these feedbacks will lead to enhanced winter increases of Arctic surface air temperature relative to the global average in response to increasing atmospheric greenhouse gases. We combine observations and global climate model simulations to examine some of the connections and feedbacks among Arctic climate variables for the present climate (control) and for a future climate in which atmospheric greenhouse gases (GHG simulation) are increasing.

For the GHG simulation, the net annual incident solar radiation at the surface decreases, but the absorbed solar radiation increases. Increased cloud cover and warmer air also cause increased downward thermal radiation at the surface. The annual increase in radiation into the ocean, however, is compensated by larger increases in sensible and latent heat fluxes out of the ocean. The seasonal cycle of surface heat fluxes is significantly enhanced. The downward flux in summer increases because there are increases in absorbed solar and thermal radiation and smaller losses of sensible and latent heat. The increased heat loss in winter is due to increased sensible and latent heat fluxes. Sea-ice cover and sea-ice thickness decrease, while cloud cover and river discharge increase. Sea-level rise occurs at a faster rate than the global average. The changes that occur exhibit spatial variability, with the changes being generally larger in coastal areas and at the ice margins. Although the well-known sea-ice albedo feedback mechanism appears to account for some of the projected temperature increase in the Arctic, hydrologic changes, such as increased water vapor and spring/summer cloud cover also play a role by increasing the flux of downward longwave radiation. A neural network approach is used to quantify relationships between these climate variables. The neural network indicates that hydrologic components of the climate system have a significant impact on the energy budget and also identifies bimodal relationships between many of the climate variables.

Raghu Murtugudde
ESSIC, University of Maryland, MD

Recent decades have seen an increasing recognition of the fact that the earth is a nonlinear system of interacting components with thresholds, hotspots, choke point, and switches leading to burgeoning interdisciplinary and multidisciplinary activities aimed at understanding the interactions among these components. The need for undergraduate and graduate education in Earth System Science has also become inevitable necessitating the development of Earth System Models of Intermediate Complexity (EMICs). ESSIC at UMD in collaboration with GFDL, WHOI, and Princeton is developing an EMIC to focus on the bio-climate feedbacks with terrestrial and marine ecosystems with an OGCM coupled to a fast atmospheric GCM. The main goal is to incorporate the most important spatial and temporal scales that are crucial for subseasonal to decadal climate variability and change while keeping the model computationally efficient enough to be run in a classroom setting. The open-source model compatible with the Earth System Modeling Framework is designed to allow interactions among the Earth System components with sufficient sophistication not only for research and education but also for Earth System predictability studies and downscaling for such applications as harmful algal blooms, human pathogens, and fisheries. This talk will focus on some examples of bio-climate interactions, coastal predictions, and climate impacts on fisheries.

D. Notz, J. Jungclaus, T. Koenigk, H. Haak, J. Marotzke
Max-Planck Institute for Meteorology, Bundesstr. 53, 20146 Hamburg, Germany (notz@dkrz.de)

We examine the sensitivity of the Atlantic Meridional Overturning Circulation (AMOC) to future variations in the Arctic fresh-water cycle caused by changes in the export of sea ice and low-salinity water, as well as by increased melting of the Greenland ice sheet. Our experiments with the coupled Max-Planck-Institute climate model ECHAM/MPI-OM show that at present, the variability of the total export of fresh water from the Arctic is mainly governed by ice export through Fram Strait. This export influences, among others, the magnitude of deep convection and sea-ice formation in the Labrador Sea. In the 21st century, the amount of total fresh-water export from the Arctic remains almost unchanged, but the export has a different geographic distribution and is mostly in the form of low-salinity water rather than sea ice. The export through the Canadian Archipelago increases, which leads in combination with locally increased P-E to a significant weakening of the deep convection in the Labrador Sea. On the contrary, export over the Barents shelf is reduced, whereas that through the Fram Strait remains almost constant. Because of the warming and freshening of the surface waters, the AMOC at 30 N declines from about 22 Sv in the 20th century to 16 Sv at the end of the 21st century. These results were obtained in IPCC experiments that did not include an interactive ice sheet model. To examine the additional impact of the predicted melting of the Greenland inland ice on the AMOC, we performed sensitivity experiments with an additional fresh-water flux around Greenland. This additional fresh-water flux was set to 0.03 Sv, as estimated from diagnosed melting rates in the coupled model runs. In one experiment, it was upscaled to 0.09 Sv to account for uncertainties connected to the predicted melting rates. The AMOC reduction for these two scenarios is 35% and 42%, respectively, compared to a weakening of 30% for the original A1B scenario. The reason for this relatively modest additional effect of the Greenland melt water is mostly that its bulk affects deep-water formation in the Labrador sea, where most of the convection has already been shut off in the standard global-warming IPCC experiment. The deep-water formation in the Greenland Sea and the overflows over the Greenland-Scotland ridge are only weakly influenced by the additional melt-water inflow. In general, the AMOC strength is more sensitive to warming in the North Atlantic than to freshening, which could point to the fact that the results may be model depending. In an outlook to future developments of the sea-ice component of the MPI-OM, we discuss briefly some shortcomings of the current thermodynamical sea-ice model and their possible impact on the discussed model results.

Response of Greenland Ice Sheet to the Global Warming Simulated by a GCM coupled by an Ice Sheet Model

Fuyuki SAITO (1), Ayako Abe-Ouchi (2,1), Tomonori Segawa (1)
1: JAMSTEC, FRCGC/JAMSTEC, Yokohama, Japan (saitofuyuki@jamstec.go.jp)
2: CCSR, Univ. of Tokyo

We compute contribution to sea level from the Greenland ice sheet under future global warming scenarios using an Atmosphere/Ocean coupled GCM and an ice sheet model. The AOGCM used in this study is the Model for Interdisciplinary Research on Climate, version 3.2 (MIROC3.2), medium (T42, 1.4x0.56 degrees) resolution, developed at the Center for Climate System Research, University of Tokyo (CCSR), National Institute for Environmental Studies (NIES) and Frontier Research Center for Global Change (FRCGC) [K-1 model developers, 2004]. The ice sheet model used in this study is Ice sheet Model for Integrated Earth system Studies (IcIES) developed at CCSR and FRCGC [Saito and Abe-Ouchi, 2004; in press]. We estimate a short-time scale response by A.D. 2100 as well as a long-time scale response after more than thousands years. Comparison is made between scenarios, coupling types, and so on. In addition, we also present several model uncertainties in response of Greenland ice sheet due to numerical procedures or parameterization schemes adopted in the models.

Hydro-thermal sensitivity of polar terrestrial climate to thermal soil representation evaluated by a 1-D model and a GCM

Kazuyuki SAITO
Frontier Research Center for Global Change (FRCGC), JAMSTEC, Yokohama, Japan (ksaito@jamstec.go.jp)

The land cryosphere, mostly underlain by perennially or seasonally frozen ground, is projected to experience a largest increase in near-surface temperatures under global warming at the end of this century. Changes in hydro-thermal regimes in those regions with degradation of permafrost, changes in depth of the active and/or seasonally-frozen layers, are anticipated to exert a large impact on local socio-economy and eco-climate systems. However, influences of those changes have potential to affect climate in other regions, and of the entire globe, through several physiochemical (e.g. albedo, freshwater discharge to the Arctic) and biological (e.g. anaerobic decomposition of tundra) pathways. Knowledge and understanding of the hydro-thermal mechanisms in the soil-freezing regions, and channels linking the local disturbances to global consequences, are essential for accurate and reliable modeling to estimate the impacts at global climate change.

Among the most fundamental of those processes are the thermal soil processes, which control the distribution of energy under the ground. Changes in thermal conductivity and capacity parameterization (e.g. consideration of unfrozen water, organic layers), and in resolved thickness and numbers of the soil layers can significantly alter the resulting hydro-thermal regimes. On the other hand, complexity and resolution need be optimized for use in global climate models (GCMs). This presentation will show preliminary evaluations on thermal regime sensitivity to different thermal soil representations by employing a 1-D physical model (MATSIRO), which computes heat and water transfer at surface and subsurface, including the effect of canopy structure, snow cover and vegetation biophysical processes. The hydro-thermal impact on the global scale will also be examined by CCSR/NIES/FRCGC version 5.7b AGCM coupled to MATSIRO conducted at a moderate horizontal resolutions (T42; typical land grid interval being about 250 km). These results will add strategic information for future improvement in polar terrestrial modeling.

Pacific Ocean inflow: Influence on catastrophic reduction of sea ice cover in the Arctic Ocean

Koji Shimada¹, Takashi Kamoshida¹, Motoyo Itoh¹, Shigeto Nishino¹, Eddy Carmack², Fiona McLaughlin², Sarah Zimmermann², and Andrey Proshutinsky³

¹Institute of Observational Research for Global Change, Japan Agency for Marine-Earth Science and Technology, Yokosuka, Japan.
² Fisheries and Oceans Canada, Institute of Ocean Sciences, Sidney, British Columbia, Canada.
³Woods Hole Oceanographic Institution, Woods Hole, Massachusetts,

The spatial pattern of recent ice reduction in the Arctic Ocean is similar to the distribution of warm Pacific Summer Water (PSW) that interflows the upper portion of halocline in the southern Canada Basin. Increases in PSW temperature in the basin are also well-correlated with the onset of sea-ice reduction that began in the late 1990s. However, increases in PSW temperature in the basin do not correlate with the temperature of upstream source water in the northeastern Bering Sea, suggesting that there is another mechanism which controls these concurrent changes in ice cover and upper ocean temperature. We propose a feedback mechanism whereby the delayed sea-ice formation in early winter, which began in 1997/1998, reduced internal ice stresses and thus allowed a more efficient coupling of anticyclonic wind forcing to the upper ocean. This, in turn, increased the flux of warm PSW into the basin and caused the catastrophic changes.

Multi-decadal variability of Atlantic Water heat transports as seen in the CCM3

Kara Sterling and Uma Bhatt
Geophysical Institute, University of Alaska Fairbanks

Changes in oceanic heat transports from the North Atlantic to the Arctic, via Atlantic Water (AW), can have widespread impacts upon Arctic climate. Using a multi-century control simulation from the National Center for Atmospheric Research (NCAR) Community Climate Systems Model version 3.0 (CCSM3), the natural multi-decadal variability (MDV) of AW is characterized.  Calculations of AW volume fluxes and heat transports into the Arctic are analyzed for the Svinøy transect, Fram Strait, and Barents Sea Opening (BSO), and compared with observations.  Warm and cold phases of AW are examined through composite analysis, and quantified with respect to their effects on Arctic climate.

The model captures several key features of AW, such as the overall circulation and depth of the AW core, but over-estimates AW temperatures by about 1 °C. AW heat anomalies can be tracked from the Svinøy transect to the Arctic interior with a timescale of 13 years, which is comparable to observations. Composites reveal a deepening (shoaling) of the AW core during warm (cold) periods. Warm (cold) periods are also characterized by greater AW transports through the BSO (Fram Strait), implying the existence of an internal ocean feedback mechanism that helps to regulate oscillations of AW between warm/cold periods.

Present Status and Future Direction of Climate Modelling in Japan

Akimasa Sumi
Center for Climate System Research,The University of Tokyo

Modelling activity for Global Warming has been accelerated by the Earth Simulator Project from 1997 to 2002, where the maximum sustain speed is 40TFLOPS. It should be noted that a project for developing a “JAPAN model” wase started together with developing a hardware. The project ise denoted to be a “Kyosei” project. That consists of three components,that is,(1) development of a high-resolution climate model(this is the first component, where CCSR,NIES and FRCGC participated, and called as K-1),(2) development of a carbon-cycle model coupled with a climate model(K-2, FRCGC), and (3) a high-resolution global atmospheric model and a regional non-hydrostatic atmospheric model(K-3,MRI/JMA). The K-1 high resolution model(MIROC 3.2) is T106L56AGCM coupled with 1/4x1/6 degrees 48 level OGCM. We developed K-1 multi-model system, where T213,T106,and T42 AGCM is available and the high-resolution and the medium resolution(1x1 degree) OGCM. Atmospheric and oceanic components can be interchanged respectively. In the K-4, MRI developed 20-km AGCM and 5km non-hydrostatic model(NHM). Results of K-1,K-2 and K-4 will be briefly presented. Detailed results will be presented by accompanied papers in this conference.

Japan has just started a project to develop the Earth Simulator follow-on, the Peta FLOPS machine developing project, where the maximum sustain speed is several Peta FLOPS. When this machine becomes available, we will develop a cloud-resolving global atmosphereic model (NICAM) and the Earth System model with full carbon cycle and dynamic vegetation model. These future plan will be briefly presented.

Climate modeling toward the next novel level: Information Transport and Humanosphere Model

Nori Tanaka
Earth Observation and Research Center, Japan Aerospace Exploration Agency, Tokyo Japan and International Arctic Research Center, University of Alaska, Fairbanks, Fairbanks, Alaska, USA

Climate model has been advanced drastically in the past 50 years as our computing capability and information technology (IT) explodes. Up to now, the climate model mainly deals with energy and material exchange on earth, which includes atmosphere, lithosphere, hydrosphere and biosphere. I try to convince you in this paper that there are additional important overlooked components in such models, which definitely needs to incorporate in the models to make them further improve the usefulness of the model as a realistic tool for better managing the earth.  One is information transport that is carried mainly by energy and material flows. Currently, the importance of those flows was judged by the amount of energy and mass. However, there are much information, which definitely regulates, modulates or changes important climate processes especially in biosphere by none or negligibly small amount of energy and mass, such as pheromones, endocrine disruptors, etc.  Without much of knowledge on this, we may encounter catastrophic crashes on ecosystems and the consequent severe impact on human population. The other one is to bring the concept of humanosphere in the climate model. Humanosphere is defined as space, which human could influence.  In the past, humanosphere is just a tinny spots on the earth but, in some sense, it is expanding in exposing power at present time. ?Soon, humanosphere will expand to include all parts of the earth, including stratosphere and deep ocean floor with uncontrollable manner, if we do business as usual. Definitely, there is an appropriate size or extent of humanosphere on the earth. Climate model should go to the direction to incorporate humanosphere concept and give us a guide to design it on the future earth.

Dynamical Understanding of the Arctic Oscillation as a Singular Eigenmode of the Global Atmosphere

H. L. Tanaka¹ and Koji Terasaki^2
1Center for Computational Sciences, University of Tsukuba, Japan, Frontier Research Center for Global Change, JAMSTEC, Japan
²Graduate School of Life and Environmental Sciences, University of Tsukuba, Japan

In this study, eigenmodes and singular modes are analyzed for a dynamical system of the atmosphere linearized about a winter basic state in order to understand the dynamics of the Arctic Oscillation (AO). Since the fluctuations of the sea-level pressure are dynamically equivalent to the barotropic component of the atmosphere, the AO is investigated in the framework of a barotropic model.

According to the result of the analysis, SVD-1 under a reasonably strong Rayleigh friction shows a structure similar to the observed AO. It is demonstrated that the singular eigenmode of the dynamical system emerges resonantly as the SVD-1 in response to the arbitrary forcing. There is no doubt that the AO may be understood as a dynamical normal mode of the global atmosphere. In reference to the result of the nonlinear simulation of the AO using the same barotropic model, we may conclude that the AO is a physical mode of a dynamical system for the global atmosphere.

Observed energetics analysis shows that the singular eigenmode is excited by the zonal-wave interactions from stationary planetary waves at the Rhines scale. The low-frequency variability associated with the AO is maintained by energy flux from the Rhines scale, which is compensated by the up-scale energy cascade from synoptic eddies in addition to the forced stationary planetary waves by topography. The external forcing such as topography explains, however, only 12% of the total variance of the AO. Therefore, AO is mostly excited by the internal nonlinear process of the interactions with the synoptic eddies.

Interaction between climate system, vegetation carbon, soil carbon, active layer thickness and fires, implementation into Earth system models

Sergey Venevsky
Hadley Centre for Climate Prediction and Research Met Office, Fitzroy Road, Exeter, EX1 3PB, UK

With temperatures projected to rise in the future, the consequences of soil thaw and permafrost melt on landscape hydrology and emissions of greenhouses gases, methane and CO2, potentially has major implications for the global carbon cycle, and thus climate change. Applications of process-oriented ecological models (LPJ DGVM, TEM etc) for scenarios of future climate change indicate carbon sequestration in Arctic tundra as increased plant production exceeds increment in decomposition. However, CH4 emissions may increase significantly in response to the soil temperature change over next century. Changes in future fire regimes and related soil and vegetation environment changes may significantly influence land-atmosphere feedbacks in circumpolar Arctic. To assess the influence of permafrost and fire regimes for vegetation structure and carbon pools/exchange the modified version of the LPJ-DGVM was run in two modes, with and without active-layer processes. Shifts of fire regime patterns in the permafrost zone caused by changes of seasonal soil moisture content dynamics were analysed for the historical period of 1901-1998. This period was chosen in order to have consistency with previous investigations of the global fire patterns, because the basic indicator of the local fire regime, fire return interval (FRI), is dependent on the averaging period. The modified LPJ DGVM was run with the historical climate data for the period 1901-1998, provided by Climate Research Unit of the University of East Anglia, in two modes with and without active-layer processes. The resulting fire return intervals (FRI) in the permafrost zone appeared to be significantly different in the both runs (see Figure 17). The FRI in western and eastern Siberia increased two to four times, especially in the continental areas, when the processes observed in the permafrost zone were included in the model. In the middle and southern taiga regions of Central Jakutia FRI has decreased four times to between 100-200 years, compared with the observed 50 to 150 years. The fire regime patterns of North American permafrost zone also improved. For instance, FRI near St. James Bay in Western Quebec decreased from more than 400 years to almost 150 years, compared with the 100 years observed. In Interior Alaska, North West Yukon, FRI shifted from 200 years to less than 100 years, while observed FRI is between 26 to 113 years. Recent simulation experiments also indicate that radiative forcing, provided by the albedo feedback may exceeds those induced by chances in carbon fluxes between the atmosphere and Arctic ecosystems. A scheme of incorporation of above mentioned processes in a family of Hadley Earth System Models (based on modification of land surface schemes, incorporated into General Circulation Models is presented.

Spatial variability of recent and projected warming in the IPCC models

John E. Walsh
International Arctic Research Center, University of Alaska Fairbanks

The global climate simulations performed for the Fourth Assessment of the Intergovernmental Panel on Global Change (IPCC) include simulations of the 20th and 21st centuries by more than a dozen different models.  External forcing for the 20th-century simulations corresponds to historical variations of greenhouse gases, while the 21st-century simulations are driven by several different greenhouse gas scenarios.  The complex spatial pattern of recent warming in the Arctic provides a natural focus for studies of the attribution of recent changes.  Additional opportunities for diagnosis and attribution arise from the strong seasonality of the observed and projected changes in the Arctic.

The outstanding features of the Arctic temperature variations over the past 50 years are the migratory “hot spots” associated with low-frequency variations of the large-scale atmospheric circulation.  Individual models show some of this behavior, as the strongest trends over the past half-century are found in different areas of the Arctic in different models.  When aggregated over all models, however, the cancellation of circulation-driven natural variations results in a pattern of modest warming (~1 ºC) over much of the Arctic, with a maximum near the sea ice margin in autumn and winter.  This pattern may be regarded as the signature of the greenhouse warming in the models.  The implication of the model results is that a 15-member ensemble of different realizations of the past 50 years of Arctic climate would show a much smoother spatial pattern, with an area-averaged-warming comparable that of the aggregate of the models’ simulations of the past half-century.

The spatial pattern of the model-projected warming intensifies in the 21st century, showing a strong signature of sea ice retreat.  This signature is similar in all three IPCC greenhouse scenarios, although it strongest in the scenario (A2) with the most rapid increase of greenhouse gas concentrations.  The prominence of the sea ice signal in the pattern of greenhouse simulations points to the importance of realistic simulations of sea ice and ice-ocean interactions in the coupled models.  The most recent observational data (2001-2005) contains indications of a similar signature associated with the extreme sea ice retreat of recent summers.  However, a comparison of the seasonal cycles of the projected and recently observed warmings shows little correspondence outside of the Arctic.  Together with the recent assessment by Serreze and Francis (2006, The Arctic amplification debate, Climatic Change), the results imply that natural variability embodied in the large-scale atmospheric circulation has dominated much of the recent local warming in the Arctic, but that the spatial and seasonal signatures of the models’ greenhouse signature are beginning to emerge in the evolution of the actual atmosphere.

Simulating the 20th century Arctic climate variability using a global coupled atmosphere-ice-ocean model 

Jia Wang
International Arctic Research Center, University of Alaska Fairbanks

Abstract
The simulations of the Arctic ice-ocean circulation using the high resolution global coupled atmosphere-ice-ocean model with 1/6x1/4 degrees and 48 vertical layers on the ‘Earth Simulator’ supercomputer are evaluated to determine the model performance, physics soundness, and its sensitivity to different process parameterizations. The model was parameterized by GM (Gent McWilliams 1990) parameterization to the north of 45N. The statistical time series of the total oceanic and ice kinetic energy and ice areas suggest that the model reaches an equilibrium without any T/S restoring or flux adjustment, and no model drifting is found. The model climatology (mean over all the model years) and variability were examined and compared with the available observations, such as ice area, temperature and salinity at certain key depths and transects. Several important physical features in the Northern Hemisphere, such as the thermohaline structure in the Arctic Ocean, Atlantic Water, meridional overturning, transports from Bering Strait, Fram Strait etc., were examined to determine physical soundness of the model. An important achievement is that the Atlantic Layer in the Arctic can be reasonably reproduced with no restoring temperature and salinity to observations. An important criterion of reproducing the Atlantic Layer variability is measured by the core (max) temperature of the layer of 500-1500m. The model reproduces reasonably the Atlantic Water core temperature in the 20th century that compares well with the observation by Polyakov et al. (2004). The model catches the 1930s-40s warming and the 1990s warming, similar to the observations. These results indicate that this coupled global model captures most important dynamic and thermodynamic processes in the Arctic Ocean. Further analyses of the model performance is underway.

Arctic Dipole Anomaly (DA) and its contribution to sea ice exports in the 20th century

Eiji Watanabe¹ and Jia Wang²

¹Center for Climate System Research, University of Tokyo
²International Arctic Research Center, University of Alaska Fairbanks

The winter dipole anomaly (DA) in the Arctic atmosphere and its contribution to sea ice export are investigated by using a high-resolution coupled general circulation model. The spatial distributions of the first two leading EOF modes of winter mean sea level pressure and geopotential height at 500 hPa north of 70oN obtained by the long-term simulation (1900-2010) are highly similar to them derived from the NCEP/NCAR reanalysis datasets (1948-2004). The first leading mode corresponds to the Arctic Oscillation (AO). The DA is defined as the second-leading mode. The AO and DA account for 66 % and 13 % of the variance, respectively.

Composite spatial patterns of sea level pressure, sea ice thickness and velocity in the extreme years when both the absolute values of PC1 and PC2 exceed 1.0 indicate that the DA plays a great important role in sea ice export from the Arctic Ocean to the Greenland Sea due to its strong meridionality. Sea ice export is highly promoted (restricted) in the positive (negative) DA phase. The dependence of sea ice export on the DA is comparable to or rather larger than that on the AO. However, whether the DA is physically independent of the AO or not has been unknown yet. Composite SLP fields suggest that the location of the most dominant anomaly in the Arctic seems to be characterized by the DA, while the sign of the anomaly is represented by the AO. In future, we should clarify the mechanism for existence of the DA.

Figure: Simulated winter mean sea level pressure anomaly regressed to PC1 and PC2 (hPa). Solid and dashed lines represent 99% and 95% significant level.

How does surface albedo feedback affect Arctic climate change?

Michael Winton
GFDL/NOAA, Michael.Winton@noaa.gov

We use simulations of the IPCC AR4 climate models to explore the role of surface albedo feedback (SAF) in anthropogenically forced Arctic climate change. Globally, the SAF is found to be the smallest and least variable of the major radiative feedbacks. The model mean value of 0.3 W/m2/K agrees with estimates from an earlier generation of climate models in spite of the incorporation of sophisticated dynamic sea ice components in the current models. The northern hemisphere accounts for 3/4 of the total feedback -- this, in turn, is about equally split between land and ocean regions.

The role of surface albedo feedback in the factor of two larger Sensitivity of Arctic temperature change relative to global is explored by comparing Arctic and global energy balances. All of the forcing and feedback factors analyzed are significantly different in the two regions and so the Arctic amplification arises from a balance of effects favoring and opposing enhanced Arctic warming. Favoring, in order of relative importance, are: longwave feedback, convergence of heat from the ocean and atmosphere and SAF. Opposing are: the non-SAF shortwave feedback and the direct CO2 forcing. Ways to improve the energy balance analysis method are discussed.

A reason for interest in the surface albedo feedback is the potential for it to change as climate change progresses. In diffusive energy balance models, a large increase in SAF as a temperature threshold is reached leads to the small ice cap instability, a rapid and difficult to reverse disappearance of polar ice. We look for the same sort of behavior in two of the AR4 models that warm to the point of eliminating Arctic sea ice year-around under 4XCO2 forcing. Neither model experiences much change in the SAF as the polar ocean becomes seasonally ice free at about -9C, but when the polar temperature reaches -5C, there is a large increase of the SAF accompanying a transition of albedo decline from late to early in the sunlit season. One of the models has a significantly larger SAF increase at this transition than the other and its ice-loss is abrupt and complete. The other model shows evidence of significant oceanic forcing of the transition and continues to have interannual appearance of ice in the Arctic ocean. The impact of the transition on polar temperature change relative to that of the Arctic and globe is also discussed.

Thermodynamic and Hydrological Impacts of Increasing Greenness in Northern High Latitudes

Jing Zhang¹, and John E.Walsh^2
1Geophysical Institute University of Alaska Fairbanks
²International Arctic Research Center, University of Alaska Fairbanks

Satellite remote sensing data indicates that greenness has been increasing in the northern high latitudes, apparently in response to the warming of recent decades. In order to identify feedbacks of this land cover change to the atmosphere, we employed the atmospheric general circulation model ARPEGE-CLIMAT to conduct a set of control and sensitivity modeling experiments. In the sensitivity experiments, we increased the greenness poleward of 60°N by 20% to mimic the manifestation of vegetation changes in the real world, and by 60% and 100% to represent potential aggressive vegetation change scenarios under global warming.  In view of the direct exposure of vegetation to sunlight during the warm seasons, we focused our study on the results from late spring to early fall. Our results revealed significant thermodynamic and hydrological impacts of the increased greenness in northern high latitudes, resulting in a warmer and wetter atmosphere. Surface and lower tropospheric air temperature showed a marked increase, with a warming of 1-2°C during much of the year when greenness is increased by 100%. Precipitation and evaporation also showed a notable increase of 10% during the summer. Snow cover decreased throughout the year, with a maximum reduction in the spring and early summer. The above changes are attributable to the following physical mechanisms: (1) increased net surface solar radiation due to a decreased surface albedo and enhanced snow-albedo feedback as a result of increased greenness; (2) intensified vegetative transpiration by the additional plant cover; and (3) reduced atmospheric stability leading to enhanced convective activity. The results imply that increased greenness is a potentially significant contributing factor to the amplified polar effects of global warming.